DOCSLIB.ORG

Clinical Trial Design As a Decision Problem

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Clinical Trial Design As a Decision Problem Special Issue Paper Received 2 July 2016, Revised 15 November 2016, Accepted 19 November 2016 Published online 13 January 2017 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/asmb.2222 Clinical trial design as a decision problem Peter Müllera*†, Yanxun Xub and Peter F. Thallc The intent of this discussion is to highlight opportunities and limitations of utility-based and decision theoretic arguments in clinical trial design. The discussion is based on a specific case study, but the arguments and principles remain valid in general. The exam- ple concerns the design of a randomized clinical trial to compare a gel sealant versus standard care for resolving air leaks after pulmonary resection. The design follows a principled approach to optimal decision making, including a probability model for the unknown distributions of time to resolution of air leaks under the two treatment arms and an explicit utility function that quan- tifies clinical preferences for alternative outcomes. As is typical for any real application, the final implementation includessome compromises from the initial principled setup. In particular, we use the formal decision problem only for the final decision, but use reasonable ad hoc decision boundaries for making interim group sequential decisions that stop the trial early. Beyond the discussion of the particular study, we review more general considerations of using a decision theoretic approach for clinical trial design and summarize some of the reasons why such approaches are not commonly used. Copyright © 2017 John Wiley & Sons, Ltd. Keywords: Bayesian decision problem; Bayes rule; nonparametric Bayes; optimal design; sequential stopping 1. Introduction We discuss opportunities and practical limitations of approaching clinical trial design as a formal decision problem. Using a case study as a running example keeps the argument focused and specific. We review a study that was set up to compare a hydrogel sealant (Progel) against standard care for patients who develop air leaks after pulmonary resection. The main features of the design are the elicitation of a utility function that quantifies clinical preferences for time to resolve the air leaks and a nonparametric Bayesian prior for the distributions of the resolution time under the two treatment arms. A Bayesian nonparametric (BNP) model is a prior for an unknown probability measure that is not restricted to a specific parametric family. Both features are important. The utility function is only meaningful if the probability model allows learning about detailed features of the event time distribution, and the nonparametric model is only needed when the decision hinges on such details. In the upcoming discussion, we focus mainly on the features of the decision problem. A complete discussion of the design and the trial appears in [1], including extensive simulations to evaluate the design’s operating characteristics (OCs) under alternative scenarios. The use of decision theoretic approaches in Bayesian clinical trial design is rare. Commonly used methods use Bayesian inference to compute posterior probabilities of clinically meaningful events or inference summaries for key parameters but then use these summaries for reasonable, but ad hoc designs. See, for example, [2] or [3] for a recent review. Beyond clinical trial design, Bayesian decision theoretic approaches are not widely used for optimal design in general. The review by [4] discusses commonly used Bayesian approaches to optimal design, focusing mainly on design problems related to learning about unknown parameters but clearly recognizing the importance of more general problems. To date, this paper remains one of the most cited and most comprehensive reviews of Bayesian optimal design. In their final discussion, Chaloner and Verdinelli write ’It is clearly helpful, in the design process, to carefully consider the reason the experiment is being done and to consider what utility should be used. [...] it would also be interesting to see alternatives constructed and explored in future research’. In the following discussion, we review one such construction which is in the spirit of what Kathryn Chaloner might have wished to see. aDepartment of Mathematics, University of Texas at Austin, USA bDepartment of Applied Mathematics and Statistics, Johns Hopkins University, USA 296 cDepartment of Biostatistics, University of Texas, M.D. Anderson Cancer Center, USA *Correspondence to: Peter Müller, Department of Mathematics, University of Texas at Austin, USA †E-mail: [email protected] Copyright © 2017 John Wiley & Sons, Ltd. Appl. Stochastic Models Bus. Ind. 2017, 33 296–301 P. MÜLLER ET AL. The discussion is restricted to the particular decision problem of clinical trial design and even further focused by fol- lowing a particular example. The arguments are meant to highlight the benefits of following a formal decision theoretic setup and also to indicate the practical limitations and challenges that are involved. 2. Decision problem 2.1. Framework We set up the design problem as a Bayesian decision problem, following a formal, principled approach. See, for example, [5] or [6] for a description of the general framework. Briefly summarized, the ingredients of a Bayesian decision problem are a probability model for all relevant unknown quantities, including data, future data, and unknown parameters of interest and a utility function that formalizes relative preferences of a decision maker under hypothetical data and assumed parameters. The optimal action, known as the Bayes rule, is the action that maximizes utility, in expectation over all unknown variables and conditional on all known variables. To be specific, let y denote the data. Often, some data may already be observed at , the time of decision making. In that case, we partition the data vector into (y y0) denoting the observed data by y0 (but we will simply use y when all data is observed). Let denote unknown parameters, and let d denote the decision. The Bayesian , , probability model usually is given as a sampling model for the data, p(y y0 ∣ d) and a prior model, p( ). Including d here in the conditioning set is a slight abuse of notation and only indicates that the sampling model can be indexed by d. In general, the prior also could be indexed by d, but this usually is not performed. Finally, let u(d,,y) denote the utility, defined as a function of a possible action d, assumed parameters , and hypothetical data y. The utility function represents the decision maker’s relative preferences for actions under an assumed truth and data. Let A denote the action set of all possible decisions under consideration. The Bayes rule is ⋆ ,, , , . d = arg max u(d y) p( y ∣ y0 d) dy d (1) d∈A ∫ ∫ ,, , , ⋆ Letting U(d ∣ y0)= u(d y) p( y ∣ y0 d) dy d , we can write d = arg max U(d ∣ y0). Here, the conditioning bar in ⋅ U( ) indicates the conditioning on y0 in the expectation. One can argue, from first principles, that a rational decision maker should act as if he or she were maximizing expected utility U(d ∣ y0) [7]. An important detail in the setup is the statement of the action set A. Usually, the set of possible actions is highly restricted, to avoid unintuitive, unreasonable, or impractical decisions. A good choice of action set avoids awkward solutions d⋆. Mathematically, a choice of a probability model and a utility function implies an optimal decision d⋆. But the mapping is very indirect through the integration and maximization in (1), and technical details in the choice of the probability model and utility function could lead to unintended solutions if care is not taken to restrict A suitably. We will give examples of this later on. 2.2. Terminal decision The design of the Progel study involves two different types of decisions. After each patient cohort is treated, and their , outcomes are observed, we decide whether to continue accrual (Sequential Stopping Decision). Let ac ∈{0 1} denote the , , continuation decision, with ac = 0 indicating continuation and ac = 1 for early stopping. We index cohorts by c = 1 … C, with a1 denoting the continuation decision after the first cohort, etc. The study includes a maximum sample size. That is,we , restrict aC = 1. Upon stopping, we decide whether or not to report Progel as superior (Terminal Decision). Let d ∈{0 1} denote the terminal decision, with d = 0 for recommending standard care versus d = 1 for Progel. In summary, the action , , set is ac ∈{0 1} and d ∈{0 1}. Ideally, all decisions should be made as Bayes rule with respect to the same underlying utility function. However, this is where the need to construct a practicable implementation that actually can be used to conduct clinical trial parts with the principled approach. For the Progel trial, only the terminal decision d is implemented as a Bayes rule. In contrast, the stopping decisions are based on a reasonable group sequential decision boundary. Details of the latter are discussed in the next section. We first discuss the terminal decision and explain the terminal decision graphically, in Figure1. Posterior inference. Index j = 1 for Progel and j = 0 for standard care, with sample sizes n1 and n0 Let yji denote the , , . time until resolution of air leaks for patient i = 1 ··· nj in treatment arm j Let Gj(y) denote the distribution of time to , , , . resolution of air leaks, j = 0 1. That is, yji ∼ Gj independently for all i = 1 ··· nj We will use a BNP model for Gj. That is, Gj itself is the unknown parameter, for which the BNP model defines a prior p(Gj). Figure 1(a) shows hypothetical posterior means for G and G , conditional on all patients up to a particular time during the trial. For some patients, actual 0 1 297 event times are recorded, while for others the time until resolution of air leaks is censored at the current time.
Load more

Recommended publications
  • Structural Equation Models and Mixture Models with Continuous Non-Normal Skewed Distributions
    Structural Equation Models And Mixture Models With Continuous Non-Normal Skewed Distributions Tihomir Asparouhov and Bengt Muth´en Mplus Web Notes: No. 19 Version 2 July 3, 2014 1 Abstract In this paper we describe a structural equation modeling framework that allows non-normal skewed distributions for the continuous observed and latent variables. This framework is based on the multivariate restricted skew t-distribution. We demonstrate the advantages of skewed structural equation modeling over standard SEM modeling and challenge the notion that structural equation models should be based only on sample means and covariances. The skewed continuous distributions are also very useful in finite mixture modeling as they prevent the formation of spurious classes formed purely to compensate for deviations in the distributions from the standard bell curve distribution. This framework is implemented in Mplus Version 7.2. 2 1 Introduction Standard structural equation models reduce data modeling down to fitting means and covariances. All other information contained in the data is ignored. In this paper, we expand the standard structural equation model framework to take into account the skewness and kurtosis of the data in addition to the means and the covariances. This new framework looks deeper into the data to yield a more informative structural equation model. There is a preconceived notion that standard structural equation models are sufficient as long as the standard errors of the parameter estimates are adjusted for failure to meet the normality assumption, but this is not correct. Even with robust estimation, the data are reduced to means and covariances. Only the standard errors of the parameter estimates extract additional information from the data.
    [Show full text]
  • On Curved Exponential Families
    U.U.D.M. Project Report 2019:10 On Curved Exponential Families Emma Angetun Examensarbete i matematik, 15 hp Handledare: Silvelyn Zwanzig Examinator: Örjan Stenflo Mars 2019 Department of Mathematics Uppsala University On Curved Exponential Families Emma Angetun March 4, 2019 Abstract This paper covers theory of inference statistical models that belongs to curved exponential families. Some of these models are the normal distribution, binomial distribution, bivariate normal distribution and the SSR model. The purpose was to evaluate the belonging properties such as sufficiency, completeness and strictly k-parametric. Where it was shown that sufficiency holds and therefore the Rao Blackwell The- orem but completeness does not so Lehmann Scheffé Theorem cannot be applied. 1 Contents 1 exponential families ...................... 3 1.1 natural parameter space ................... 6 2 curved exponential families ................. 8 2.1 Natural parameter space of curved families . 15 3 Sufficiency ............................ 15 4 Completeness .......................... 18 5 the rao-blackwell and lehmann-scheffé theorems .. 20 2 1 exponential families Statistical inference is concerned with looking for a way to use the informa- tion in observations x from the sample space X to get information about the partly unknown distribution of the random variable X. Essentially one want to find a function called statistic that describe the data without loss of im- portant information. The exponential family is in probability and statistics a class of probability measures which can be written in a certain form. The exponential family is a useful tool because if a conclusion can be made that the given statistical model or sample distribution belongs to the class then one can thereon apply the general framework that the class gives.
    [Show full text]
  • Maximum Likelihood Characterization of Distributions
    Bernoulli 20(2), 2014, 775–802 DOI: 10.3150/13-BEJ506 Maximum likelihood characterization of distributions MITIA DUERINCKX1,*, CHRISTOPHE LEY1,** and YVIK SWAN2 1D´epartement de Math´ematique, Universit´elibre de Bruxelles, Campus Plaine CP 210, Boule- vard du triomphe, B-1050 Bruxelles, Belgique. * ** E-mail: [email protected]; [email protected] 2 Universit´edu Luxembourg, Campus Kirchberg, UR en math´ematiques, 6, rue Richard Couden- hove-Kalergi, L-1359 Luxembourg, Grand-Duch´ede Luxembourg. E-mail: [email protected] A famous characterization theorem due to C.F. Gauss states that the maximum likelihood es- timator (MLE) of the parameter in a location family is the sample mean for all samples of all sample sizes if and only if the family is Gaussian. There exist many extensions of this re- sult in diverse directions, most of them focussing on location and scale families. In this paper, we propose a unified treatment of this literature by providing general MLE characterization theorems for one-parameter group families (with particular attention on location and scale pa- rameters). In doing so, we provide tools for determining whether or not a given such family is MLE-characterizable, and, in case it is, we define the fundamental concept of minimal necessary sample size at which a given characterization holds. Many of the cornerstone references on this topic are retrieved and discussed in the light of our findings, and several new characterization theorems are provided. Of particular interest is that one part of our work, namely the intro- duction of so-called equivalence classes for MLE characterizations, is a modernized version of Daniel Bernoulli’s viewpoint on maximum likelihood estimation.
    [Show full text]
  • DISTRIBUTIONS for ACTUARIES David Bahnemann
    CAS MONOGRAPH SERIES NUMBER 2 DISTRIBUTIONS FOR ACTUARIES David Bahnemann CASUALTY ACTUARIAL SOCIETY This monograph contains a brief exposition of the standard probability distributions and their applications by property/casualty actuaries. The focus is on the use of parametric distributions fitted to empirical claim data to solve standard actuarial problems, such as creation of increased limit factors, pricing of deductibles, and evaluating the effect of aggregate limits. A native of Minnesota, David Bahnemann studied mathematics and statistics at the University of Minnesota and at Stanford University. After teaching mathematics at Northwest Missouri State University for 18 years, he joined the actuarial department at the St. Paul Companies in St. Paul, Minnesota. For the next 25 years he provided actuarial support to several excess and surplus lines underwriting departments. While at the St. Paul (later known as St. Paul Travelers, and then Travelers) he was involved in large-account and program pricing. During this time he also created several computer- based pricing tools for use by both underwriters and actuaries. During retirement he divides his time between White Bear Lake and Burntside Lake in Minnesota. DISTRIBUTIONS FOR ACTUARIES David Bahnemann Casualty Actuarial Society 4350 North Fairfax Drive, Suite 250 Arlington, Virginia 22203 www.casact.org (703) 276-3100 Distributions for Actuaries By David Bahnemann Copyright 2015 by the Casualty Actuarial Society. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or trans- mitted, in any form or by any means. Electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of the publisher.
    [Show full text]
  • (Severini 1.2) (I) Defn: a Parametric Family {P
    Chapter 2: Parametric families of distributions 2.1 Exponential families (Severini 1.2) (i) Defn: A parametric family {Pθ; θ ∈ Θ} with densities w.r.t. some σ-finite measure of the form Pk f(x; θ) = c(θ)h(x) exp( j=1 πj(θ)tj(x)) − ∞ < x < ∞. (ii) Examples: Binomial, Poisson, Gamma, Chi-squared, Normal, Beta, Negative Bionomial, Geometric ... (iii) NOT: Cauchy, t-dsns, Uniform, any where support depends on θ. (iv) Note c(θ) depends on θ through {πj(θ)}. Defn: (π1, ..., πk) is natural parametrization. – defined only up to linear combinations. We assume vector π is of full rank – no linear relationship between the πj. (v) Natural parameter space R Pk Π = {π : 0 < h(x) exp( j=1 πj(θ)tj(x))dx < ∞} Lemma: Π is convex. (vi) Thm: For any integrable φ, any π0 in interior of R Pk Π, then φ(x)h(x) exp( j=1 πj(θ)tj(x))dx is cts at π0, has derivatives of all orders at π0, and can get derivatives by differentiating under the integral sign. Cor: With φ ≡ 1, c(π) is cts, diffble, etc. (vii) Moments: T = (t1(X), ....., tk(X)) Pk Note log f(x; π) = log(h(x)) + log c(π) + j=1 πjtj(x). ∂f ∂ log f Also ∂π = ∂π f. R ∂ Differentiating f(x; π)dx = 1 gives E(T ) = −∂π (log c(π)) ∂2 Differentiating again gives var(T ) = −∂π2 (log c(π)) 1 2.2 Transformation Group families (Severini 1.3) (i) Groups G of transformations on <: Contains identity, closed under inverses, and composition.
    [Show full text]
  • Compact Parametric Models for Efficient Sequential Decision
    Compact Parametric Models for Efficient Sequential Decision Making in High-dimensional, Uncertain Domains by Emma Patricia Brunskill Submitted to the Department of Electrical Engineering and Computer Science in partial fulfillment of the requirements for the degree of Doctor of Philosophy at the MASSACHUSETTS INSTITUTE OF TECHNOLOGY June 2009 c Massachusetts Institute of Technology 2009. All rights reserved. Author................................................ ............. Department of Electrical Engineering and Computer Science May 21, 2009 Certifiedby.......................................... ............... Nicholas Roy Assistant Professor Thesis Supervisor Acceptedby........................................... .............. Terry P.Orlando Chairman, Department Committee on Graduate Theses Compact Parametric Models for Efficient Sequential Decision Making in High-dimensional, Uncertain Domains by Emma Patricia Brunskill Submitted to the Department of Electrical Engineering and Computer Science on May 21, 2009, in partial fulfillment of the requirements for the degree of Doctor of Philosophy Abstract Within artificial intelligence and robotics there is considerable interest in how a single agent can autonomously make sequential decisions in large, high-dimensional, uncertain domains. This thesis presents decision-making algorithms for maximizing the expcted sum of future rewards in two types of large, high-dimensional, uncertain situations: when the agent knows its current state but does not have a model of the world dynamics within a
    [Show full text]
  • Chapter 1 the Likelihood
    Chapter 1 The Likelihood In this chapter we review some results that you may have came across previously. We define the likelihood and construct the likelihood in slightly non-standard situations. We derive properties associated with the likelihood, such as the Cr´amer-Rao bound and sufficiency. Finally we review properties of the exponential family which are an important parametric class of distributions with some elegant properties. 1.1 The likelihood function Suppose x = X is a realized version of the random vector X = X .Supposethe { i} { i} density f is unknown, however, it is known that the true density belongs to the density class .Foreachdensityin , f (x) specifies how the density changes over the sample F F X space of X.RegionsinthesamplespacewherefX (x)is“large”pointtoeventswhichare more likely than regions where fX (x) is “small”. However, we have in our hand x and our objective is to determine which distribution the observation x may have come from. In this case, it is useful to turn the story around. For a given realisation x and each f 2F one evaluates f (x). This “measures” the likelihood of a particular density in based X F on a realisation x. The term likelihood was first coined by Fisher. In most applications, we restrict the class of densities to a “parametric” class. That F is = f(x; ✓); ✓ ⇥ ,wheretheformofthedensityf(x; )isknownbutthefinite F { 2 } · dimensional parameter ✓ is unknown. Since the aim is to make decisions about ✓ based on a realisation x we often write L(✓; x)=f(x; ✓)whichwecallthelikelihood.For convenience, we will often work with the log-likelihood (✓; x)=logf(x; ✓).
    [Show full text]
  • Lecture 7, Review Data Reduction 1. Sufficient Statistics
    Lecture 7, Review Data Reduction We should think about the following questions carefully before the "simplification" process: • Is there any loss of information due to summarization? • How to compare the amount of information about θ in the original data X and in 푇(푿)? • Is it sufficient to consider only the "reduced data" T? 1. Sufficient Statistics A statistic T is called sufficient if the conditional distribution of X given T is free of θ (that is, the conditional is a completely known distribution). Example. Toss a coin n times, and the probability of head is an unknown parameter θ. Let T = the total number of heads. Is T sufficient for θ? Sufficiency Principle If T is sufficient, the "extra information" carried by X is worthless as long as θ is concerned. It is then only natural to consider inference procedures which do not use this extra irrelevant information. This leads to the Sufficiency Principle : Any inference procedure should depend on the data only through sufficient statistics. 1 Definition: Sufficient Statistic (in terms of Conditional Probability) (discrete case): For any x and t, if the conditional pdf of X given T: 푃(퐗 = 퐱, 푇(퐗) = 푡) 푃(퐗 = 퐱) 푃(퐗 = 퐱|푇(퐗) = 푡) = = 푃(푇(퐗) = 푡) 푃(푇(퐗) = 푡) does not depend on θ then we say 푇(퐗) is a sufficient statistic for θ. Sufficient Statistic, the general definition (for both discrete and continuous variables): Let the pdf of data X is 푓(퐱; 휃) and the pdf of T be 푞(푡; 휃).
    [Show full text]
  • Hypothesis Testing Econ 270B.Pdf
    Key Definitions: Sufficient, Complete, and Ancillary Statistics. Basu’s Theorem. 1. Sufficient Statistics1: (Intuitively, a sufficient statistics are those statistics that in some sense contain all the information aboutθ ) A statistic T(X) is called sufficient for θ if the conditional distribution of the data X given T(X) = t does not depend on θ (i.e. P(X = x | T(X) = t) does not depend on θ ). How do we show T(X) is a sufficient statistic? (Note 1 pg 9) Factorization Theorem: In a regular parametric model {}f (.,θ ) :θ ∈ Θ , a statistic T(X)with range R is sufficient forθ iff there exists a function g :T ×Θ → ℜ and a function h defined on ℵ such that f (x,θ ) = g(T(x),θ )h(x) ∀ x ∈ℵ and θ ∈ Θ 2 Sufficient Statistics in Exponential Families: ⎡ k ⎤ ⎡ k ⎤ p(x,θ ) = h(x) exp⎢ η j (θ )T j (x) − B(θ )⎥ or px(,θθηθ )= hxC () ()exp⎢ jj () T () x⎥ ÆT (X ) = (T1 (X ),...,Tk (X )) a sufficient stat. for θ ⎢∑ ⎥ ⎢∑ ⎥ ⎣ j=1 ⎦ ⎣ j=1 ⎦ 2. Ancillary Statistics: (Intuitively, ancillary statistics are those statistics that are in some sense uninformative aboutθ .) A statistic S(X) is ancillary for a parameter θ if the distribution of S(X) does not depend onθ . It is first order ancillary if E(S(X)) does not depend onθ . ANCILLARY STATISTICS ARE IRRELEVENT FOR INFERENCE ABOUT θ Æits realized values don’t help us pin downθ since the realized values don’t depend on θ . Note: To show ancillarity, use the definition! Intuition for The Need for Completeness: For some sufficient statistics, T(X), if r(T(X)) is ancillary for some function r Æ T(X) contains some extraneous information about θ .
    [Show full text]
  • A Flexible Parametric Family for the Modeling and Simulation of Yield Distributions
    Journal of Agricultural and Applied Economics, 42,2(May 2010):303–319 Ó 2010 Southern Agricultural Economics Association A Flexible Parametric Family for the Modeling and Simulation of Yield Distributions Octavio A. Ramirez, Tanya U. McDonald, and Carlos E. Carpio The distributions currently used to model and simulate crop yields are unable to accom- modate a substantial subset of the theoretically feasible mean-variance-skewness-kurtosis (MVSK) hyperspace. Because these first four central moments are key determinants of shape, the available distributions might not be capable of adequately modeling all yield distributions that could be encountered in practice. This study introduces a system of distributions that can span the entire MVSK space and assesses its potential to serve as a more comprehensive parametric crop yield model, improving the breadth of distributional choices available to researchers and the likelihood of formulating proper parametric models. Key Words: risk analysis, parametric methods, yield distributions, yield modeling and simulation, yield nonnormality JEL Classifications: C15, C16, C46, C63 Agricultural economists have long recognized Shonkwiler (1993), Ramirez, Moss, and Boggess that the choice of an appropriate probability (1994), Coble et al. (1996), and Ramirez (1997). distribution to represent crop yields is critical These authors have provided strong statistical for an accurate measurement of the risks as- evidence of nonnormality and heteroskedasticity sociated with crop production. Anderson in crop yield distributions—specifically, the ex- (1974) first emphasized the importance of ac- istence of kurtosis and negative skewness in counting for nonnormality in crop yield distri- a variety of cases. The possibility of positive butions for the purpose of economic risk anal- skewness has been documented as well (Ramı´rez, ysis.
    [Show full text]
  • Lecture Notes on Advanced Statistical Theory1
    Lecture Notes on Advanced Statistical Theory1 Ryan Martin Department of Statistics North Carolina State University www4.stat.ncsu.edu/~rmartin January 3, 2017 1These notes were written to supplement the lectures for the Stat 511 course given by the author at the University of Illinois Chicago. The accompanying textbook for the course is Keener's Theoretical Statistics, Springer, 2010, and is referred to frequently though out the notes. The author makes no guarantees that these notes are free of typos or other, more serious errors. Feel free to contact the author if you have any questions or comments. Contents 1 Introduction and Preparations 4 1.1 Introduction . .4 1.2 Mathematical preliminaries . .5 1.2.1 Measure and integration . .5 1.2.2 Basic group theory . 10 1.2.3 Convex sets and functions . 11 1.3 Probability . 12 1.3.1 Measure-theoretic formulation . 12 1.3.2 Conditional distributions . 14 1.3.3 Jensen's inequality . 16 1.3.4 A concentration inequality . 17 1.3.5 The \fundamental theorem of statistics" . 19 1.3.6 Parametric families of distributions . 21 1.4 Conceptual preliminaries . 22 1.4.1 Ingredients of a statistical inference problem . 22 1.4.2 Reasoning from sample to population . 23 1.5 Exercises . 25 2 Exponential Families, Sufficiency, and Information 29 2.1 Introduction . 29 2.2 Exponential families of distributions . 30 2.3 Sufficient statistics . 33 2.3.1 Definition and the factorization theorem . 33 2.3.2 Minimal sufficient statistics . 36 2.3.3 Ancillary and complete statistics . 38 2.4 Fisher information .
    [Show full text]
  • 1. Introduction
    Virtual Laboratories > 4. Special Distributions > 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 1. Introduction In this chapter, we will study a number of parametric families of distributions that have special importance in probability. In some cases, a distribution may be important because it occurs as the limit of other distributions. In some cases, a parametric family may be important because it can be used to model a wide variety of random phenomena. In turn, this is usually the case because the family has a rich collection of densities with a small number of parameters (usually 1 or 2). As a general philosophical principal, we try to model a random process with as few parameters as possible; this is sometimes referred to as the principal of parsimony of parameters. In turn, this is a special case of Ockham's razor, named in honor of William of Ockham, the principle that states that one should use the simplest model that adequately describes a given phenomenon. There are several other parametric families of distributions that are studied elsewhere in this project, because the natural home for these distributions are various random processes. These include The binomial distribution The negative binomial distribution The multinomial distribution The hypergeometric distribution The multivariate hypergeometric distribution The Pólya distribution The Poisson distribution Before we begin our study of special parametric families of distributions, we will study two general parametric families. Many of the special parametric families studied in this chapter belong to one or both of these general families. Location-Scale Families 1.
    [Show full text]